pfSense

2. In the case of kern.ipc.nmbclusters the default is too high for low end platforms such as uFW / SG-1100.
(eg. on SG-1100 kern.ipc.nmbclusters defaults to 1000000 which is 2GB of buffer memory - the SG-1100 only has 1GB of RAM - fortunately FreeBSD does not pre-allocate memory)
This inhibits advanced users attempting to configure a more performant yet conservative configuration (eg. avoid memory exhaustion from DoS attacks)

The code automatically removes the default values for those on 22.01 on ARM (64-bit and 32-bit) so they should no longer have a defined value, leaving the OS to determine the value. Letting the user override that could be problematic on that hardware.

The value is not touched on other platforms, the user is free to define whatever they feel is necessary.

The default is high, but tuning it depends on their hardware and environment. There isn't any good method or guidance to figure it out automatically since it has to be set before the hardware is initialized to avoid some problems (e.g. on systems with high end NICs with lots of queues, the hardware can fail to initialize if the value is too low). It can also increase under load but the exact values again vary based on the hardware, network environment, configuration, load, etc. Better to set it too high and not need it than to set it too low and hit kernel panics. Though FreeBSD is better here now than it has been in the past.

Running 22.05.a.20220402.0600 on the 1100, kern.ipc.nmbclusters is not present in /boot/loader.conf or system tunables

Jordan Greene wrote in #note-4:

Running 22.05.a.20220402.0600 on the 1100, kern.ipc.nmbclusters is not present in /boot/loader.conf or system tunables

This is the correct behavior for the 1100 appliance (pfsense-plus firmware), see /etc/inc/pfsense-utils.inc#L1152